Risk Assessment
LinkRisk Analysis
Every domain linking to your site is classified into one of five risk bands. Our analysis goes far beyond basic spam scores — it's powered by the largest disavow dataset outside of Google.
How LinkRisk Works
LinkRisk isn't a single metric or a simple threshold check. It's a multi-layered scoring algorithm built on over 150 individual rulesets, each designed to detect a specific pattern, behaviour, or risk signal associated with manipulative or low-quality linking domains.
Every linking domain enters the algorithm with a neutral baseline score of LR500. From there, the score is nudged up or down as each ruleset evaluates the domain. Positive trust signals — like consistent editorial linking patterns, strong domain age, clean backlink profiles, and topical authority — push the score higher. Negative signals — spam patterns, link selling indicators, PBN footprints, sudden link velocity spikes, or presence in our disavow dataset — pull it down.
The rulesets aren't weighted equally. Some carry more influence than others. Appearing across multiple independent disavow files, for example, is a heavily weighted negative signal — it represents real-world decisions by experienced SEOs who concluded that domain was harmful. Conversely, a domain with decades of consistent editorial content and natural link acquisition patterns receives strong positive reinforcement.
The final score determines which of the five risk bands a domain falls into. Because the algorithm uses over 150 rulesets working in combination, it's extremely difficult for manipulative domains to game the system — a domain might pass a handful of checks, but the cumulative weight of dozens of subtle signals paints an accurate picture.
This approach was developed over two decades of hands-on link analysis and refined against our proprietary disavow dataset — the largest outside of Google. It's not theory. It's pattern recognition at scale, built by people who've been doing this since before Penguin existed.
Example LinkRisk Breakdown
Here's what a typical LinkRisk distribution looks like in one of our reports. Each segment shows the proportion of linking domains in each risk band.
Clean, authoritative domains
Generally trustworthy with minor imperfections
Mixed signals, worth monitoring
Patterns consistent with link schemes
High-risk, recommend disavow
Example data from a real client audit (anonymised)
The Five Risk Bands
Each linking domain is categorised based on dozens of signals. Here's what each band means and what signals drive the classification.
Good
Clean, authoritative domains with strong trust signals, natural link profiles, and no history of spam or manipulation.
Key Signals
- High domain authority
- Natural link velocity
- Clean backlink profile
- Established domain age (3+ years)
- No history of penalties
Low Risk
Generally trustworthy domains with minor imperfections. These links are safe and contribute positively to your profile.
Key Signals
- Moderate authority
- Some minor spam signals
- Mostly natural link patterns
- Newer but growing domains
- No penalty history
Neutral
Domains with mixed signals that neither help nor harm your profile significantly. Worth monitoring but not urgent action required.
Key Signals
- Low-moderate authority
- Mixed link patterns
- Some commercial intent in outbound links
- Limited editorial standards
- No clear spam but no strong trust either
Suspect
Domains showing patterns consistent with link schemes, thin content farms, or PBN-like characteristics. These should be reviewed for potential disavow.
Key Signals
- Abnormal link velocity spikes
- High outbound link ratio
- Thin or auto-generated content
- Multiple domains on same IP range
- Patterns matching known link networks
Bad
High-risk domains with strong spam indicators. Links from these domains should be disavowed to protect your site from algorithmic or manual penalties.
Key Signals
- Known spam or hacked domains
- Present in our disavow dataset from multiple sources
- Extremely high outbound link counts
- Malware or phishing history
- Previously penalised by Google
How We Assess Risk
Our LinkRisk classification uses five complementary analysis methods, each contributing to the final risk band assignment.
Disavow Dataset Cross-Reference
Every domain is checked against our proprietary disavow dataset — the largest outside of Google. If a domain appears in multiple disavow files from different sites, that's a strong negative signal. We weight this heavily because it represents real-world decisions by experienced SEOs who determined these domains were harmful.
Spam Pattern Detection
Our algorithms analyse over 40 spam indicators including: outbound link density, content quality signals, ad-to-content ratio, anchor text patterns in outbound links, and cross-referencing with known link selling networks. Domains are flagged when multiple spam patterns co-occur.
Link Velocity Analysis
Sudden spikes in a domain's outbound linking activity often indicate paid link campaigns or compromised sites. We track linking velocity over time and flag domains that show unnatural growth patterns inconsistent with organic editorial linking.
Network Topology
We map the relationships between linking domains — shared hosting, shared ownership patterns, and interlinking. Clusters of domains that heavily interlink and share infrastructure are a hallmark of private blog networks (PBNs).
Historical Penalty Data
Using our dataset, we identify domains that have been associated with Google penalties — both algorithmic (Penguin) and manual actions. Domains with penalty history carry elevated risk even if current metrics appear clean.
How risky is your link profile?
Want to get a free report on one of your sites and find out which links are helping and which are not?
Free Report